Modeling dust emission in the Magellanic Clouds with Spitzer and Herschel

Dust modeling is crucial to infer dust properties and budget for galaxy studies. However, there are systematic disparities between dust grain models that result in corresponding systematic differences in the inferred dust properties of galaxies. Quantifying these systematics requires a consistent fitting analysis. We compare the output dust parameters and assess the differences between two dust grain models, Compiegne et al (2011), and THEMIS (Jones et al 2013, Kohler et al 2015). In this study, we use a single fitting method applied to all the models to extract a coherent and unique statistical analysis. We fit the models to the dust emission seen by Spitzer and Herschel in the Small and Large Magellanic Clouds(SMC and LMC). The observations cover the infrared (IR) spectrum from a few microns to the sub-millimeter range. For each fitted pixel, we calculate the full n-D likelihood, based on the method described in Gordon et al (2014). The free parameters are both environmental ($U$, the interstellar radiation field strength; $\alpha_\mathrm{ISRF}$, power-law coefficient for a multi-U environment; $\Omega^*$, the starlightstrength) and intrinsic to the model ($Y_\mathrm{i}$: abundances of the grain species $i$; $\alpha_\mathrm{sCM20}$, coefficient in the small carbon grain size distribution). Fractional residuals of 5 different sets of parameters show that fitting THEMIS brings a more accurate reproduction of the observations than the Compiegne model. However, independent variations of the dust species show strong model-dependencies. We find that the abundance of silicates can only be constrained to an upper-limit and the silicate/carbon ratio is different than that seen in our Galaxy. In the LMC, our fits result in dust masses slightly lower than those found in literature , by a factor lower than 2. In the SMC, we find dust masses in agreement with previous studies.